Training-free Subject-Enhanced Attention Guidance for Compositional Text-to-image Generation

TL;DR

This paper introduces a training-free, subject-driven guidance method for compositional text-to-image generation that improves attribute binding and subject fidelity, especially in zero-shot scenarios.

Contribution

It proposes a novel training-free inference guidance framework that enhances attention maps for better compositional generation without additional training.

Findings

Improves subject fidelity and attribute binding in generated images.

Achieves strong zero-shot compositional generation performance.

Introduces GroundingScore metric for evaluating subject alignment.

Abstract

Existing subject-driven text-to-image generation models suffer from tedious fine-tuning steps and struggle to maintain both text-image alignment and subject fidelity. For generating compositional subjects, it often encounters problems such as object missing and attribute mixing, where some subjects in the input prompt are not generated or their attributes are incorrectly combined. To address these limitations, we propose a subject-driven generation framework and introduce training-free guidance to intervene in the generative process during inference time. This approach strengthens the attention map, allowing for precise attribute binding and feature injection for each subject. Notably, our method exhibits exceptional zero-shot generation ability, especially in the challenging task of compositional generation. Furthermore, we propose a novel metric GroundingScore to evaluate subject alignment thoroughly. The obtained quantitative results serve as compelling evidence showcasing the effectiveness of our proposed method. The code will be released soon.